tape emulation

[wenzel.hellgren]

http://www.cirrus.com/en/pubs/whitePaper/aes04.pdf

[soundguy]

thats a nice read, although Im at a loss to see how it backs your statement about anti aliasing.

have you done any listening comparisons with replacing the input opamp to your converter with a transformer?

dave

[wenzel.hellgren]

thats a nice read, although Im at a loss to see how it backs your statement about anti aliasing.

have you done any listening comparisons with replacing the input opamp to your converter with a transformer?

dave

It wasn't meant to back any statement, it's just a nice read. Did you actually read it?

because it does discus a buffer stage that is resistive, without amplification (if I read that right, it says it attenuates the signal, doesn't say it amplifies it) and has the anti-alias filter in it.

the PDF seems to be protected so I can't copy the relevant parts.

check part 3 : Input Buffer Design. Possibly part 3.3 too, but I don't want to read the whole thing again to check.



how do you replace the input stage if the entire converter is on a single chip? I thought most devices used single chip converters.

so, no, I have not done any listening tests. I've never seen an A/D converter that has a transformer.

[Family Hoof]

how do you replace the input stage if the entire converter is on a single chip? I thought most devices used single chip converters.

Well, I was under the impression that good converter companies, while they're stuck with whatever digital chips the semi people put out, do design their own surrounding circuitry and analog front end (based on prefab monolithic op amps, of course!) ...but I wouldn't be surprised if the shameless cheap stuff is simply stamping a name on an all-in-one IC box. Like I said, I have yet to investigate the practical considerations of converer design, only basic concepts, so I don't know everything that's out there, parts-wise. Can anyone else shed some more light on this for us? Any digital designers hanging out at the TOMB?

[soundguy]

I did actually skim it over.

If the entire converter and all associated active stages in the box are on one chip then you are shit outta luck. This example is the prime reason to shy away from integrated circuits in professional audio- if you dont like the input buffer they stuck on the same chip as the converter, you are stuck with it. I really dont know if this is the norm or not though.

I spoke at length with a designer about this issue about six months ago working towards a custom box before I eventually gave up on the idea. I got the impression then that there are two actual digital covnverter chips that do the A/D conversion that are popular among the big 5 manufacturers, I was led to believe that everyone was basically using the same converter chip, just implementing it with their own twist and own buffers. I never opened any boxes to bullshit test that statement, its just what I was told and Im passing it along, take it for what it is. So long as the input buffer is a seperate active stage, as it should be, then you can stick anything that will do the job in there that you want from a IC opmap to a discrete opamp to a transformer.

The paper shows a 6dB loss provided by the input buffer. This is easily achieved by the right transformer so no biggie.

In regards to anti aliasing, what I get from reading that paper is the input buffer is spec'd to provide a given CMRR and impedance in order for the filter before the converter chip to function properly not that the filter is actually in the input buffer chip itself, which I dont believe it to be. I am most certainly not an expert with this stuff by any stretch of the imagination so I could very well be wrong.

What I am is a dude who heard a shitty alesis converter improved by simply replacing the buffer chip with a transformer. There are some other guys online who have done this as well. Id put my friends box up next to a lavry any day.

Is there a reason not to use a transformer? For the digital programming guy who is probably designing converters, of course. Transformers spec for shit on paper compared to some chips. There's lots of reasons to use one however, if the device is being designed by a recording engineer instead of a numbers dude.

When you take the average converter box, that chip isnt doing anything beyond isolation. Its functioning in a similar way as it does in a mic pre or a console output section, etc. Compare a rev E 1176 with a rev F. With the F, they replaced the input attenuator network/transformer that was in the E with a chip and associated circuitry. This change probably resulted in more than $100 savings for urei, and thats probably 1970's $100 too. Its the same crap with these converters. When you get into the "good" converters you can bet that they are using the best chips available but I still think if these companies TRULY were obsessed with the way these units sound if they really didnt want to use a transformer that they'd at least design a discrete input that had some depth to it which so far I have not heard from a chip of any kind.

Is this tweakerville? Well, yeah, kinda. Unless you are doing everything discrete its probably up for debate wether or not you are gonna hear the difference if you mixed on say an SSL and then have to run through a converter with an IC front end. In a case like that, this whole discussion is likely lost on that scenario. I would venture that most people recording today are using IC's all over the place so this is really not something I expect the world to stand up and take notice about. Manufacturers are building stuff to the lowest common denomonator their customers will tolerate and thats what we are stuck with. IF you are lucky enough to be in a situation however where you are able to make a record without using a single IC anywhere, not in a mic pre, not in an EQ, not in a tape deck, not in a console, imagine the frustration of taking all that effort and cramming it between the micro layers of some dip, your entire record, as the very last thing your audio sees before digital conversion. Well aint that a kick in the balls?

The reason I believe we dont have converters like this is not because its impossible but because manufacturers are cheap or because these boxes are not designed by engineers or both. The other side of the coin is that there are people out there, designers, that really believe in the IC. Not everyone is making rock records. Stereo classical recording is gonna sound just fine through a top of the line converter. People make all kinds of music. Unfortunately, people make only one kind of converter. If you dont fit with their program, you shit outta luck. We, as a community, that is professional engineers, are a bunch of sheep who dont confront manufacturers with our needs. This is what happens when recording engineers are just knob turners. They cant communicate with designers on even a 'fo technical level, so they get stuck with whatever designers give them. Corners keep getting cut, users mouths keep staying shut, its a bad cycle. Designers have pacified the real needs of engineers with nice looking knobs and leds. Used to be stuff had to sound good to be good. Now stuff just has to sound good compared to whats available on the marketplace in order for it to be good. You could put shit on rye, wheat or multigrain, its still a shit sandwich.

so no, you dont see a digital converter with a passive front end. Thats fucking horrifying if you ask me. Its a big giant kick in the face to anyone who wants to use a computer but is serious about the way their shit sounds. It can be done, it just isnt being done because there isnt a vocal demand for it. My friend modified his converter, its been stable for a while now, I dont think what Im asking for is an outlandish request, unfortunately for me Im half retarded so cant design anything like this so I just keep saying this hoping the right smart guys are listening. If you make a truly awesome converter with a passive front end I think lots of people would dig it if given the opportunity to listen to it and compare for themselves.

dave

[Family Hoof]

This is what happens when recording engineers are just knob turners. They cant communicate with designers on even a 'fo technical level, so they get stuck with whatever designers give them. Corners keep getting cut, users mouths keep staying shut, its a bad cycle.

AMEN!!!

[I'm Painting Again]

many people have modified converters by removing the active buffers in cheap converters and replacing them with transformers. Its very easy to do. Unfortunately, if you dont want to rip apart your gear and modify it, there arent manufacturers that offer this type of product prepackaged.

dave

I think the Lavry blue and gold converters have transformer balanced buffers..

they sound awesome I use the blue ones..they sound like what you put into them..and they soft limit and compress..have different settings for saturation..they have meters that can get 1db accuracy and they are calabratable in terms of setting a point below full scale to be your +4 "0"..

I didn't really like the sound of digital recording before them..everything else I had been using was horribly "wrong" sounding to me..now I'm pretty content..

[soundguy]

Id be suprised if the lavry have trannsformer front ends. Dan lavry is definitely the supergenius, but I think he's a fan of the dip. They are really good converters, I used them on my last record, definitely the best thing going that Ive heard so far.

dave

[I'm Painting Again]

Id be suprised if the lavry have trannsformer front ends. Dan lavry is definitely the supergenius, but I think he's a fan of the dip. They are really good converters, I used them on my last record, definitely the best thing going that Ive heard so far.

dave

i dont know for sure but i remeber in the manual something about transformer buffers..i dont even really know what we are talking about here though..

[Family Hoof]

I just read over the Lavry Blue manual and it looks like there are transformers in some of the digital paths, specifically the AES/EBU input, but they're not used for analog buffering the A/D section... but I'd love to be proven wrong.

[Family Hoof]

Came across this paper in my daily reading and thought it'd be worth posting here. Shows two examples of (presumably) popular input buffers for A/D and discusses common mode rejection. That corp says their chip can yeild a minimum CMRR@60 Hz in excess of 70 dB (CMMR = common mode rejection ratio), and this is apparently better than competing chips. Unfortunately 70dB is pretty pathetic compared to a good transformer. Here's one example.